Retail investors have reportedly already placed more than $100 billion in orders for SpaceX stock
Retail investors have reportedly already placed more than $100 billion in orders for SpaceX stock—more than the entire market capitalization of most Fortune 100 companies—for a company that lost $4.9 billion in fiscal 2025 and has burned through $41.3 billion since its founding in 2002. The S-1 filing tells them explicitly that SpaceX will pay no dividends "in the foreseeable future" and will retain future earnings to finance the growth of its business.
This is the deal being pitched as the largest IPO in history, targeting a valuation in the range of $1.5 to $2 trillion that would put SpaceX among the five most valuable publicly traded companies on Earth. Retail participation in the offering has reportedly hit somewhere in the low-20% range, according to reports—triple the normal 5-10% retail allocation—meaning everyday investors are lining up to own a piece of a company whose only demonstrably profitable business is being systematically strip-mined to fund two ventures that show no credible path to positive unit economics.
For every dollar of AI revenue, SpaceX lost two dollars.
The brutal segmental math tells the story management would prefer you not notice. Starlink, the satellite internet service, generated $11.4 billion in revenue and $4.4 billion in operating income in fiscal 2025. That is 61% of total company revenue and, critically, the only segment that made money. The legacy Space segment—the rocket launch business that has been operating since 2002—remains unprofitable after nearly a quarter century, and in 2025 alone spent $3 billion on Starship research and development. The AI segment posted approximately $6.4 billion in operating losses on $3.2 billion in revenue. Read that again: for every dollar of AI revenue, SpaceX lost two dollars.
The capital allocation picture is even starker. Of the $20.7 billion SpaceX spent on capex in fiscal 2025, a full 61%—$12.6 billion—went to the AI segment. Starlink, the actual profit engine, received a fraction of that. The company is explicitly telling investors that the business model is to take cash generated by a working connectivity service and light it on fire building GPU clusters and rockets that may never generate a positive return, while retail shareholders foot the bill and Elon Musk's 10-vote Class B shares ensure he retains unilateral control over every capital allocation decision in perpetuity.
Management's TAM slide is where the IPO pitch crosses from aggressive into fantasy. SpaceX claims a $28.5 trillion total addressable market, which it calls "the largest actionable total addressable market in human history." Of that $28.5 trillion, exactly $26.5 trillion—93% of the entire opportunity—is assigned to AI, including a breathtaking $22.7 trillion attributed to "enterprise applications" of artificial intelligence. The proven Space segment gets $370 billion. Starlink connectivity, the only part of the business that actually prints money today, is assigned $1.6 trillion. The valuation story rests almost entirely on AI economics that do not yet exist and that the S-1's own numbers show are currently destroying capital at a rate that would terrify any venture investor operating under normal fiduciary constraints.
The Starlink unit economics are deteriorating, not improving. Average revenue per user collapsed from $99 per month in 2023 to $66 per month by March 2026, a 33% decline in just over two years, even as subscriber acquisition costs remain elevated and growth is decelerating. Starlink added 2.1 million subscribers between 2023 and 2024, then 4.5 million between 2024 and 2025, but the S-1 shows the company at 10.3 million total subscribers by March 2026—implying growth is flattening as the service saturates its highest-value customer segments and pushes into lower-ARPU international markets. One analyst quoted in the evidence notes that without Starship achieving a dramatic further reduction in launch costs—on top of Falcon 9's already 85% cost advantage over legacy systems—Starlink's infrastructure may not be profitable serving more than 25 to 30 million households.
That is a hard ceiling on the connectivity business, and it arrives well before Starlink reaches the scale needed to justify a trillion-dollar segment valuation.
Starship is the linchpin holding this entire structure together, and Starship does not work yet. The rocket suffered three consecutive test failures in 2025 before partial successes in August and October 2025, and as of the May 20 S-1 filing had never delivered a payload to orbit. The company states it expects Starship to begin payload delivery in the second half of 2026 and to start deploying next-generation V3 Starlink satellites in 2027—timelines that leave zero room for the kind of setbacks that have defined the program's history. The S-1 lists Starship delays as the top risk factor to the entire business, noting that failure to achieve "sufficient launch cadence, reusability, and capability" would "materially and adversely affect" the ability to execute the growth strategy.
That growth strategy includes not just satellite deployment but also the orbital AI data centers that are supposed to justify the $26.5 trillion AI TAM. If Starship does not reach full operational capability at the cadence and cost structure management projects, the AI infrastructure story collapses, Starlink's next-generation capacity never materializes, and investors are left holding equity in an unprofitable rocket company subsidized by a connectivity business whose unit economics are eroding.
The governance structure ensures that none of this matters to Elon Musk's control. The dual-class share setup gives Musk's Class B shares 10 votes each, while public Class A shares get one vote. The S-1 explicitly states that SpaceX will be a "controlled company" under Nasdaq rules, meaning it is not required to maintain a majority of independent directors or independent compensation and nominating committees. Public shareholders will have zero ability to influence capital allocation, block dilutive acquisitions, or prevent Musk from reallocating Starlink profits to Mars and AI regardless of returns.
Retail investors are being asked to fund moonshots they cannot vote on, in a company that will not pay dividends, controlled by a man whose other roles—CEO of Tesla, former head of a government cost-cutting initiative—are themselves listed as material risk factors because his "actions and statements, whether or not directly relating to us, may draw significant public attention and scrutiny" and could negatively impact the stock price.
The AI segment's losses are accelerating, not narrowing. The $6.4 billion operating loss in fiscal 2025 came despite $3.2 billion in revenue, and the first quarter of 2026 showed a $4.3 billion net loss for the entire company. SpaceX is burning cash faster than at any point in its history, and the S-1 discloses total long-term debt of $29.1 billion as of March 2026, alongside that $41.3 billion accumulated deficit. The company took out a bridge loan in March 2026 and amended its revolving credit facility in May 2026 to increase capacity—signs that liquidity, not long-term value creation, is driving the IPO timing. Meanwhile, the AI segment consumed 61% of total capex in 2025 while growing revenue at just 22%, a figure well below the reported growth rates at frontier AI labs like OpenAI and Anthropic, raising the question of whether xAI's models are competitive at all.
SpaceX is pitching retail investors on a company where proven cash flows subsidize speculative bets indefinitely, where the only profitable business is cannibalizing itself through ARPU decline and decelerating growth, where the capital-intensive projects that dominate spending have missed every major timeline and burned $41 billion cumulative without reaching profitability, and where governance protections are zero because one man's super-voting shares let him redirect all proceeds to Mars colonies and AI inference whether or not those ventures ever pay off. The valuation in the range of $1.5 to $2 trillion is not a multiple of earnings—the company has no earnings—but a call option on Elon Musk's vision, priced as if the vision were already real. Venture capitalists would never accept this risk-return profile at this valuation.
Retail investors are being asked to accept it in the low-20% range of the deal, betting $100 billion that rockets and AI will somehow generate returns that connectivity alone cannot, in a structure designed to ensure they have no voice if the bet goes wrong.